Techniques have been developed to support user interaction with shapes in a user interface. These techniques support creation of a shape in an image, specify selection of an object within an image, perform a cropping operation, and so forth. Conventional techniques that are utilized to support this interaction, however, often lack intuitiveness and are not natural to a casual user, thereby resulting in user frustration and inefficiencies.
For example, conventional techniques to manipulate vector shapes are performed using vector points and handles, aka Bezier paths. Thus, this interaction is not performed with the vector shape itself in these conventional techniques, but rather with vector points and handles that are positioned away from the shape outline in order to make a change to the shape. Consequently, casual users are typically forced to guess how to input a change by making guesses as to which vector points and handles provide a desired effect, thereby resulting in an arduous trial-and-error process to achieve a desired result.